The history of the Indo-Europeans Tandy Warnow The University of Texas at Austin Questions about Indo-European (IE) • How did the IE family of languages evolve? • Where is the IE homeland? • When did Proto-IE “end”? • What was life like for the speakers of protoIndo-European (PIE)? The Kurgan Expansion • Date of PIE ~4000 BCE. • Map of Indo-European migrations from ca. 4000 to 1000 BC according to the Kurgan model • From http://indo-european.eu/wiki The Anatolian hypothesis (from wikipedia.org) Date for PIE ~7000 BCE Estimating the date and homeland of the proto-Indo-Europeans • Step 1: Estimate the phylogeny • Step 2: Reconstruct words for protoIndo-European (and for intermediate proto-languages) • Step 3: Use archaeological evidence to constrain dates and geographic locations of the proto-languages DNA Sequence Evolution -3 mil yrs AAGACTT AAGGCCT AGGGCAT AGGGCAT TAGCCCT TAGCCCA -2 mil yrs TGGACTT TAGACTT AGCACTT AGCACAA AGCGCTT -1 mil yrs today U AGGGCAT V W TAGCCCA X TAGACTT Y TGCACAA X U Y V W TGCGCTT Standard Markov models of biomolecular sequence evolution • Sequences evolve just with substitutions • There are a finite number of states (four for DNA and RNA, 20 for aminoacids) • Sites (i.e., positions) evolve identically and independently, and have “rates of evolution” that are drawn from a common distribution (typically gamma) • Numerical parameters describe the probability of substitutions of each type on each edge of the tree Rates-across-sites B D A C B A D C • Dates at nodes are only identifiable under ratesacross-sites models with simple distributions, and also requires an approximate lexical clock. Violating the rates-across-sites assumption • The tree is fixed, but do not just scale up and down. • Dates are not identifiable. C A B D A B C D Linguistic character evolution • Homoplasy is much less frequent: most changes result in a new state (and hence there is an unbounded number of possible states). • The rates-across-sites assumption is unrealistic • The lexical clock is known to be false • Borrowing between languages occurs, but can often be detected. These properties are very different from models for molecular sequence evolution. Phylogeny estimation requires different techniques. Dating nodes requires both an approximate lexical clock and also the rates-across-sites assumption. Neither is likely to be true. Historical Linguistic Data • A character is a function that maps a set of languages, L, to a set of states. • Three kinds of characters: – Phonological (sound changes) – Lexical (meanings based on a wordlist) – Morphological (especially inflectional) Sound changes • • Many sound changes are natural, and should not be used for phylogenetic reconstruction. Others are bizarre, or are composed of a sequence of simple sound changes. These are useful for subgrouping purposes. Example: Grimm’s Law. 1. 2. 3. Proto-Indo-European voiceless stops change into voiceless fricatives. Proto-Indo-European voiced stops become voiceless stops. Proto-Indo-European voiced aspirated stops become voiced fricatives. Homoplasy-free evolution • When a character changes state, it changes to a new state not in the tree • In other words, there is no homoplasy (character reversal or parallel evolution) • First inferred for weird innovations in phonological characters and morphological characters in the 19th century, and used to establish all the major subgroups within IndoEuropean. 0 0 1 0 0 0 0 1 1 Lexical characters can also evolve without homoplasy • For every cognate class, the nodes of the tree in that class should form a connected subset - as long as there is no undetected borrowing nor parallel semantic shift. 1 1 1 0 0 0 1 1 2 Phylogeny estimation • Linguists estimate the phylogeny through intensive analysis of a relatively small amount of data – a few hundred lexical items, plus – a small number of morphological, grammatical, and phonological features • All data preprocessed for homology assessment and cognate judgments • All “homoplasy” (parallel evolution, back mutation, or borrowing) must be explained and linguistically believable Our (RWT) Data • Ringe & Taylor (2002) – 259 lexical – 13 morphological – 22 phonological • These data have cognate judgments estimated by Ringe and Taylor, and vetted by other IndoEuropeanists. (Alternate encodings were tested, and mostly did not change the reconstruction.) • Polymorphic characters, and characters known to evolve in parallel, were removed. Our methods/models • Ringe & Warnow “Almost Perfect Phylogeny”: most characters evolve without homoplasy under a no-common-mechanism assumption (various publications since 1995) • Ringe, Warnow, & Nakhleh “Perfect Phylogenetic Network”: extends APP model to allow for borrowing, but assumes homoplasy-free evolution for all characters (Language, 2005) • Warnow, Evans, Ringe & Nakhleh “Extended Markov model”: parameterizes PPN and allows for homoplasy provided that homoplastic states can be identified from the data. Under this model, trees and some networks are identifiable, and likelihood on a tree can be calculated in linear time (Cambridge University Press, 2006) • Ongoing work: incorporating unidentified homoplasy and polymorphism (two or more words for a single meaning) First analysis: “Weighted Maximum Compatibility” • Input: set L of languages described by characters • Output: Tree with leaves labelled by L, such that the number of homoplasy-free (compatible) characters is maximized (while requiring that certain of the morphological and phonological characters be compatible). • NP-hard. The WMC Tree dates are approximate 95% of the characters are compatible Modelling borrowing: Networks and Trees within Networks “Perfect Phylogenetic Network” (all characters compatible) What about PIE homeland and date? • Linguists have “reconstructed” words for ‘wool’, ‘horse’, ‘thill’ (harness pole), and ‘yoke’, for Proto-Indo-European, and for ‘wheel’ for the ancestor of the “core” (IE minus Anatolian and Tocharian). • Archaeological evidence (positive and negative) for these objects used to constrain the date and location for proto-IE to be after the “secondary products revolution”, and somewhere with horses (wild or domesticated). • Combination of evidence supports the date for PIE within 30005500 BCE (some would say 3500-4500 BCE), and location not Anatolia, thus ruling out the Anatolian hypothesis. Acknowledgements • Financial Support: The David and Lucile Packard Foundation, the National Science Foundation, The Program for Evolutionary Dynamics at Harvard, The Radcliffe Institute for Advanced Studies, and the Institute for Cellular and Molecular Biology at UT-Austin. • Collaborators: Don Ringe (Penn), Steve Evans (Berkeley), and Luay Nakhleh (Rice) • Thanks also to Don Ringe (Penn), Craig Melchert (UCLA), and Johanna Nichols (Berkeley) for discussions related to the date and homeland for PIE • Please see http://www.cs.rice.edu/~nakhleh/CPHL for papers and data For more information • Please see http://www.cs.rice.edu/~nakhleh/CPHL (the Computational Phylogenetics for Historical Linguistics web site) for data and papers How old is PIE? (1) Words for 'yoke' and 'draw, pull (on sledge)' reconstruct to PIE, hence PIE dispersed after the development of animal traction. (2) Words for 'wool' reconstruct to PIE, hence PIE dispersed after the development of woolly sheep. (Ancestral sheep and goats have short hair -- unspinnable, unfeltable.) (3) A verb for 'milk (an animal)' reconstructs to PIE, hence PIE dispersed after the "secondary products revolution". (4) Words for 'wheel', 'thill' (harness pole), and 'convey (in a vehicle) reconstruct to at least core IE and maybe all PIE, hence PIE dispersed after (or not too long before) the development of wheeled transport. How old is PIE? (1) Words for 'yoke' and 'draw, pull (on sledge)' reconstruct to PIE, hence PIE dispersed after the development of animal traction. northern Mesopotamia, c. 4000 BCE spread from Mesopotamia c. 3000 BCE Darden, Bill J. 2001. On the question of the Anatolian origin of Indo-Hittite. In Robert Drews, ed., Greater Anatolia and The Indo-Hittite Language Family, 184-228. Washington, DC: Institute for the Study of Man. Sherratt, Andrew. 1981. Plough and pastoralism: Aspects of the secondary product revolution. In I. Hodder, G. Isaac and G. Hammond, eds., Pattern of the Past: Studies in Honour of David Clarke, 261-205. Cambridge: Cambridge University Press. How old is PIE? (2) Words for 'wool' reconstruct to PIE, hence PIE dispersed after the development of woolly sheep. (Ancestral sheep and goats have short hair -- unspinnable, unfeltable.) woolly sheep: eastern Iran, after 7000 BCE (maybe) wool: Sumeria, North Caucasus steppe after 4000 BCE Barber, E. J. W. 1991. Prehistoric Textiles: The Development of Cloth in the Neolithic and Bronze Ages. Princeton: Princeton University Press. Darden, Bill J. 2001. On the question of the Anatolian origin of Indo-Hittite. In Robert Drews, ed., Greater Anatolia and The Indo-Hittite Language Family, 184-228. Washington, DC: Institute for the Study of Man. Shishlina, N. I., O. V. Orfinskaja and V. P. Golikov. 2003. Bronze Age textiles from the North Caucasus: New evidence of fourth millennium BC fibres and fabrics. Oxford Journal of Archaeology 22.331-344. How old is PIE? (3) A verb for 'milk (an animal)' reconstructs to PIE, hence PIE dispersed after the "secondary products revolution". Darden, Bill J. 2001. On the question of the Anatolian origin of Indo-Hittite. In Robert Drews, ed., Greater Anatolia and The Indo-Hittite Language Family, 184-228. Washington, DC: Institute for the Study of Man. Sherratt, Andrew. 1981. Plough and pastoralism: Aspects of the secondary product revolution. In I. Hodder, G. Isaac and G. Hammond, eds., Pattern of the Past: Studies in Honour of David Clarke, 261-205. Cambridge: Cambridge University Press. How old is PIE? (4) Words for 'wheel', 'thill' (harness pole), and 'convey (in a vehicle)' reconstruct to at least core IE and maybe all PIE, hence PIE dispersed after (or not long before) the development of wheeled transport. c. 4000-3500 BCE in or near today's Ukraine, Romania Anthony, David W. 2007. The Horse, the Wheel, and Language: How Bronze Age Riders From the Eurasian Steppes Shaped the Modern World. Princeton, NJ: Princeton University Press. Darden, Bill J. 2001. On the question of the Anatolian origin of Indo-Hittite. In Robert Drews, ed., Greater Anatolia and The Indo-Hittite Language Family, 184-228. Washington, DC: Institute for the Study of Man. Parpola, Asko. Proto-Indo-European speakers of the Late Tripolye culture as the inventors of wheeled vehicles: Linguistic and archaeological considerations of the PIE homeland problem. In Karlene Jones-Bley, Martin E. Huld, Angela Della Volpe and Miriam Robbins Dexter, eds., Proceedings of the 19th Annual UCLA IndoEuropean Conference, 1-59. Washington, DC: Institute for the Study of Man. How old is PIE? Couldn't these words have been borrowed into the IE daughter branches millennia after the PIE dispersal? NO! Words borrowed separately into distant languages would look very different, as with medieval Arabic loans into European languages: Spanish algodon * química (reshaped!) * French coton * chemie * English cotton (< French!) † chemistry (reshaped!) † German Baumwolle (coinage!) † Chemie (from French!) † Russian xlopok (lit. 'fluff': coinage!) ximija (via Greek!) * Can't even reconstruct Proto-Romance! † Can't even reconstruct Proto-Germanic! Extended Markov model • Each character evolves down the tree. • There are two types of states: those that can arise more than once, and those that can only arise once. We also know which type each state is. • Characters evolve independently but not identically, nor in a rates-across-sites fashion. • Essentially this is a linguistic version of the nocommon-mechanism model, but allowing for an infinite number of states. Initial results • Under very mild conditions (substitution probabilities bounded away from 1 and 0), the model tree is identifiable - even without identically distributed sites. • Fast, statistically consistent, methods exist for reconstructing the tree (and the network, under some conditions). • Maximum Likelihood and Bayesian analyses are also feasible, since likelihood calculations can be done in linear time.